University of Groningen
Isotopic characterisation of CO2 sources during regional pollution events using isotopic and
radiocarbon analysis
Zondervan, A; Meijer, HAJ
Published in:
Tellus. Series B: Chemical and Physical Meteorology
DOI:
10.1034/j.1600-0889.1996.00013.x
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
1996
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Zondervan, A., & Meijer, HAJ. (1996). Isotopic characterisation of CO2 sources during regional pollution
events using isotopic and radiocarbon analysis.
Tellus. Series B: Chemical and Physical Meteorology
,
48
(4), 601-612. https://doi.org/10.1034/j.1600-0889.1996.00013.x
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.
More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-
amendment.
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Tellus (1996), 48B, 601-612
Printed in UK - all rights reserved
Copyright
©
M
unksgaard, 1996
TELLUS
ISSN 0280-6509
Isotopic characterisation of
C02
sources during regional
pollution events using isotopic and radiocarbon analysis
By ALBERT ZONDERVAN* AND HARRO A. J. MEIJER!, Centrum voor IsotopenOnderzoek,
Groningen University, 9747E AG Groningen, The Netherlands
(Manuscript received 2 October 1995;in final form 22 May 1996)
ABSTRACT
At the station Kollumerwaard (The Netherlands), for monitoring tracers in the troposphere,
air is sampled in 16containers for off-line 13C,180 and 14Cisotopic analysis of C02. The timing
of the sampling is chosen such that CO
2
variations correlating with pollutants like CO and
CH
4
are optimally covered. The 14Cmeasurements enable us to discriminate between biospheric
and fossil fuel contributions to atmospheric CO
2
, The analysis of one series sampled on
23 November 1994 resolves the increased CO
2
mixing ratio into a purely biospheric component
with a
δ13C
of (- 22.2
±
1.5)‰0,and a fossil component of up to 35 ppm with a
δ
13
C
of
(- 34.1
±
1.6)‰0.Another series, recorded on 2 and 3 February 1995, shows a nearby emission
of fossil CO
2
, methane and carbon monoxide, most likely due to the flaring of natural gas.
Both events clearly indicate the importance of natural gas consumption in or in the vicinity of
Holland. These experimental values can be compared with estimates of CO
2
emissions from
combustion of fossil fuels and the corresponding
δ13C
values. The results for 180 show the
pronounced difference in behaviour between the °and C isotopes in atmospheric CO
2
, due
to the fast isotopic exchange processes with (plant, soil or ocean) water. As a side result, the
method produces the ratio CO: fossil CO
2
, a direct measure for combustion quality on a
regional scale.
1. Introduction
Isotopic analysis of carbon dioxide in the lower
atmosphere has resulted in a better understanding
ofits exchange with the biosphere and the role of
anthropogenic activity. Studies of global carbon
fluxes involve models which are strongly con-
strained by observations of the carbon isotope
ratios. Since the pioneering work of Keeling
(Keeling, 1958; Keeling, 1961; Keeling et aI., 1979),
numerous studies have been performed on atmo-
spheric CO
2
and its isotopes on various regional
and time scales (Keeling et a!., 1989; Francey et a!.,
1990; Nakazawa et a!., 1993; Conway et a!., 1994;
1Corresponding author. Meijer@phys.rug.nl.
*Present address: Institute for Geological and Nuclear
SciencesLtd., PO Box 31312, Lower Hutt, New Zealand.
Levin et a!., 1995; Francey et a!., 1995; Keeling
et a!., 1995). Diurnal and seasonal cycles of the
C02 mixing ratio and the
13CI12C
ratio give
insight in the magnitude of the exchange on a
regional and a global scale, largely thanks to the
relatively large differences between the isotopic
signatures of the reservoirs.
In
a continuous effort
to improve the description of reality with model
calculations, grids are sized down and less domin-
ant components are taken into account wherever
sensible and possible. This goes hand in hand with
geographically more refined observations of more
(carbon containing) components.
The most obvious example is that of anthropo-
genic CO
2
, most of which is produced from com-
bustion of fossil fuels. Its role is paramount to
"balancing the carbon budget" between present
ocean, atmospheric and biospheric reservoirs
Tellus48B (1996), 4
602
A. ZONDERVAN AND H. A. 1. MEIJER
(Ciais et aI., 1995). Consumption levels of fossil
fuel, both globally and seasonally averaged, can
be calculated, e.g., from import and industrial
records (Marland and Boden, 1993). However, it
is already much more complicated to determine
seasonal and regional variations in fossil fuel (type)
use (Rotty, 1987). Since the different fossil fuel
types have different 13C/12Cratios the temporal
and seasonal behaviour of the 13C/12Cratio of
fossil fuel CO
2
can only be estimated (Keeling,
1973, Tans, 1981).
In
an effort to supply regional
fossil fuel data a database containing annual CO
2
emissions, flux and
δ13C
by country has been set
up (CDIAC, 1991; Andres et aI., 1993; Boden
et al. 1995).
From an experimental point of view, fossil fuel
CO
2
is best detectable by 14Cmeasurements, since
fossil fuel contains no 14C. Furthermore, 14C02
has been and still is monitored on many sites on
earth (Nydal and Lovseth 1983; Manning et aI.,
1990; Meijer et a!., 1994; Levin et aI., 1995), and
the so-called Suess-effect has been studied on
several of these sites (Levin et aI., 1980; Levin
et a!., 1989). Combining 13C/12C and 14C/12C
analysis of atmospheric CO
2
in principle enables
one to validate experimentally the data on fossil
fuel emissions based on the production informa-
tion mentioned above.
A 2nd example is the rôle of land ecosystems in
the exchange of CO
2
with the atmosphere.
Analysis of C1800 is particularly suitable for this
purpose as CO
2
readily exchanges its oxygen
isotopes with water in leaf tissues and soils (Friedli
et aI., 1987; Francey and Tans 1987; Farquhar
et a!., 1993; Ciais et a!., 1996; Ehleringer et a!.,
1993).
Another possible use of isotopic characteris-
ation of tracers is finger printing regional sources
of pollutants. Although the chemistry of ozone
formation in the mixing layer is theoretically
largely understood, modelling and prediction
remains heavily dependent on fine-scale transport
observations of pollutants over rural and urban
areas. Most important precursors to 03 are CO,
CH
4
and
Nax,
of which only methane has import-
ant sources besides combustion, such as microbial
activity and leakage from natural gas distribution
systems. Again, measurement of all three carbon
isotopes enables a separation of components
derived from fossil fuel. Such information would
be valuable for the assessment of anthropogenic
activity, such as the quality of combustion by
industry and vehicles.
In this paper, we present an experimental
method to identify the fossil and biospheric com-
ponents of the total CO
2
mixing ratio by their
unique isotopic signatures in a local/regional
atmosphere. In Section 2 the footing of this
method is given. Its full implementation in the
operation of a station in a rural area of the
Netherlands is explained in Section
3.
The power
and constraints of the method are illustrated in
Section 4 with the analysis of the first series of
measurements at this station.
In
Section 5 we
discuss possible improvements and the applicabil-
ity of this method to the three studies mentioned
above. Conclusions are drawn in Section
6.
2. Experimental method
The basis for the separation of the measured
total-C0
2
mixing ratio into a biospheric, a fossil
and a background component are the large differ-
ences in carbon isotopic signatures. Fossil CO
2
contains no 14C,whereas the 14C/12Cratio in the
biosphere remains close to that of the global
troposphere due to the continuous and relatively
rapid exchange between the two reservoirs.
Consequently, a measured 14C/12C below the
background value directly yields the fraction of
fossil CO
2
. Finally, the differences between the
13C/12Cratios of all three components allow a
complete separation of the measured total-C02
into the three components if sufficient information
is available. A combination with 180/160 measure-
ments may produce additional information on the
way atmospheric CO
2
interacts with the biosphere
and with the oceans.
Assuming that all background parameters are
known, a single measurement (one sample) con-
tains just enough information to assign isotopic
ratios to the biospheric and fossil components.
However, reliable results can only be obtained by
regression analysis of a series of measurements
with sufficient variability in the observables. Also,
the isotopic signatures of the individual compon-
ents must remain constant. These conditions
require the observation of a changing transport of
air from the (polluting) source region during inter-
vals in the order of hours to days.
The variables involved are:
Tellus 48B (1996), 4
ISOTOPIC CHARACTERISATION OF C02
603
- T the
C02
volume mixing ratio expressed
in ppm and on a WMO mole fraction
scale;
-b
13
C
the per mille deviation of the
13C/12C
ratio from the VPDB primary standard
and on the Vienna-PDB (VPDB) scale,
(Coplen, 1995; Allison et aI., 1995);
_b14C
the per mille deviation of the
14C/
12
C
ratio from the standardised value of the
pre-bomb atmosphere in 1950 (Stuiver
and Pollach, 1977);
_15180
the per mille deviation of the
180rO
ratio from the VPDB primary standard
and on the VPDB-C0
2
scale without
normalisation to SLAP (Coplen, 1995;
Allison et aI., 1995).
The three reservoirs are distinguished by indices
b,
f,
and a, that stand for Biospheric, Fossil, and
background Atmosphere, respectively. For each of
the isotopic ratios a budget equation applies:
Ti
=
Ta
+
Tb,i
+
Tf,i'
(1)
b13Ci
·Ti
=
b13Ca •
Ta
+
b
13
C
b
· 11"i
+
b13Cf ·Tf.i'
(2)
(1+
b14Ci)·Ti
=
(1 + b
14
C
a
)·
Ta
+(1 +
b
14
C
b
)· Tb,i,
(3)
b
18
0
i
·Ti
=
b
18
0
a
•
11"i
+
b
18
0f ·Tf,i'
(4)
where the index
i
refers to a specific sample in the
series.Measured quantities appear only left of the
equal signs. Eqs. (2), (3) and (4) are good approxi-
mations of the exact budget equations: the linearis-
ation errors are negligible. The absence of
14C
in
fossil
CO
2
has been applied to eq. (3).
Eqs.
(1)-(3)
reflect two basic assumptions in
our method. (i) During the interval of measure-
ment the parameters of the background atmo-
sphere remain constant. This is a reasonable
assumption for a time scale of a few days or less,
as the seasonal variation and the effect that air
masses mix in, originating from different latitudes,
are negligible over such intervals. (ii) The regional
biosphere is treated as a single reservoir, adding
to or taking away
CO
2
from the atmosphere with
variable amounts, while the carbon isotopic ratios
associated with this biospheric
CO
2
remain con-
stant. For carbon this assumption is valid thanks
to the high temporal resolution used in this work:
the diurnal cycle of uptake and release of
CO
2
by
plants and soils is fully resolved, and can in good
approximation be described as exchange of carbon
between two reservoirs (atmosphere and bio-
sphere) with constant isotopic composition over
this short time interval. (compare Dirks and
Goudriaan, 1995),
Assuming that plant material (B) is only in
contact with the background atmosphere (A) and
that
14C
reservoir effects can be neglected, i.e. no
time-lag of carbon isotopic ratios by storage in B,
then B is in isotopic equilibrium with A. To
account for isotopic fractionation between these
two reservoirs the quantity
Δ14C
is introduced:
Δ
14
C
c
= Δ
14
C
a
.
(5)
These quantities differ from the parameters
b
14
C,
used in equation (3), by a correction term describ-
ing the fact that
14CI12C
fractionation is 2 x that
of the
13CI12C
ratio, relative to a general reference
value for plant material of
b
13
C
=-
25%0
(Stuiver
and Pollach, 1977):
[
0.975
]2
I+Δ14C=(I+b14C)
l+b
13
C '
(6)
For a series of measurements these eqs. make
up an over-determined set, which can be solved
by maximum likelihood regression techniques
such as minimum least-squares fitting.
Eq, (4) for
15
18
0
is deliberately chosen analogous
to
b13C.
However, while assumption (i) and (ii)
are justified for carbon isotopes for time scales of
a few days or less, these assumptions will not hold
for oxygen. The equilibration process between
oxygen isotopes of atmospheric carbon dioxide
and of (ocean) water is rapid, can proceed without
any net flux of
CO
2
and depends on many met-
eorological variables (Keeling, 1958; Keeling,
1961; Bottinga and Craig, 1969; Francey and Tans,
1987, Farquhar et aI., 1993; Ciais et aI., 1996), The
rapid equilibration process for
15
18
0
with either
ocean surface water or plant and soil water makes
it impossible to define a "background air" com-
position. Large differences in
15
18
0
will exist
between oceanic and land air mass parcels, even
in the case that neither of them has experienced
any anthropogenic influence, A change of influence
of this process on the sampled air immediately
modifies in eq. (4) the background parameter
b
18
0
a
during the time-series. Similarly, assumption
(ii) cannot readily be used for exchange of oxygen
isotopes between a terrestrial ecosystem and the
atmosphere because it does not take the influence
of regional humidity on evapotranspiration and
Tellus 48B (1996), 4
604
A. ZONDERVAN AND H. A. 1. MEIJER
thereby on the
15180
of the leaf water into account
(Francey and Tans, 1987). Therefore, it is expected
that analysis of
15180
information according to
eq. (4) does not add any significance to the separa-
tion of the fossil and biospheric components from
the total-C02 signal. On the other hand, the
technique presented here is potentially very power-
ful for research into atmosphere-biosphere inter-
action using
15
18
0,
since the parallel analysis of
the carbon and oxygen isotopes enables one to
observe the specific differences in the way they
exchange between the reservoirs.
3. Experimental procedure
The monitoring station is jointly operated by
three Dutch research institutes: the Electricity
Companies joint Research Institute KEMA,
the national Institute of Public Health and
Environmental Research (RIVM), and the
Netherlands Organisation for Applied Scientific
Research (TNO). It has been set up in 1991 to
monitor radioactivity and several tracers associ-
ated with air pollution and ozone production. It
is located in the Kollumerwaard polder at 53.3°N
6.3°E, a rural area close to the Waddensea. The
city of Groningen, which has circa 200,000inhabit-
ants, is located 30 km to the SE, and forms the
largest industrialised area within a 130 km radius.
Station Kollumerwaard is surrounded by farm-
land and some woodland. There is hardly any
cattle nearby.
We have installed an air sampler system utilising
the station's existing infrastructure. From a height
of 8 m air is sucked down through a central duct,
to which each system has a connection. In the
sampler system the undried air is delivered to a
series of 16 containers which are sequentially
flushed. A 10 litre volume of the glass containers
corresponds to 0.15 mmol C02 at STP, which is
more than sufficient for an accurate isotopic ana-
lysis. A 1.5h flushing period for each volume
results in a coverage of the past 24 h. The flow of
5 l/min is maintained by an oil-free membrane
pump on the exit side. Tubes are made of either
stainless steel or Pyrex glassware. Valves are modi-
fied versions of Louwers Hapert stopcocks with
Viton O-rings. A dust filter on the entrance side
removes large dust particles. The operation of the
sampler is controlled by a Pc. Apart from system
diagnostics and operation of the containers' valves,
the PC continuously receives the latest CO
2
, CH4,
CO mixing ratios and basic meteo data via a local
network. The gas concentrations are measured by
a Chrompack FID gas chromatograph optimally
tuned for methane. The GC is calibrated by meas-
urement of four reference gas mixtures, one by
one at 3-h intervals (Van den Beld and Veldkamp,
1995). For CO
2
the calibration error is 2 ppm,
while the noise is below 0.5 ppm.
Sequential filling of the air containers can be
interrupted or stopped in two ways: by remote
control via a modem or by a user-defined trigger
in the PC control program. In the frame of the
present work, a time series of CO
2
is considered
to be of special interest for isotopic analysis only
if it contains sizeable variations (> 10%) that
correlate with carbon monoxide and, less import-
antly, with methane. Such an event is likely to
correspond with air downwind from industrialised
or urbanised areas and not solely with the diurnal
cycle of photosynthesis and respiration of plant
material. Data from an anemometer can be used
for a crude selection of the origin of the air and
to reject slow-mixing conditions. Because of the
combination of a simple sampler with real-time
diagnostics of data, we call this mode of operation
"event trapping".
Once the sequencing has been stopped, the
containers are taken to the laboratory where
carbon dioxide is extracted cryogenically. While
the air is recirculated (≈21/min), the bulk of the
water vapour is removed by a H
2
0 trap (dry ice
alcohol slush, -75°C) during the 1st 15min. Then,
the C02 trap is immersed in liquid air (-180°C)
during one h. Finally the trapped CO
2
is collected
in a sample flask immersed in liquid air, while the
trap (after thawing) is kept at dry ice temperature
to prevent possible water traces to be collected in
the flask along with the C02, This procedure
results in a high extraction efficiency and a negli-
gible fractionation. Testing with artificial N2-C02
mixtures showed a CO
2
extraction yield of
>99.9% and no change in its isotopic composi-
tions relative to the original mixture. No special
precautions are taken to exclude atmospheric
N
2
0, °
2
,CH4 and Ar from the extracted fraction.
The b
13
C and
15
18
0
of the CO
2
are measured
with a VG-Sira 9 Mass Spectrometer (MS). The
accuracy
(1-0-)
of a single stable-isotope determina-
tion, including the MS calibration error, is estim-
Tellus 48B (1996), 4